Safety-critical Policy Iteration Algorithm for Control under Model Uncertainty

Safety is an important aim in designing safe-critical systems. To design such systems, many policy iterative algorithms are introduced to find safe optimal controllers. Due the fact that most practical finding accurate information from system rather impossible, a new online training method presented this paper perform reinforcement learning based algorithm using real data instead of identifying dynamics. Also, impact model uncertainty examined on control Lyapunov functions (CLF) and barrier (CBF) dynamic limitations. The Sum Square program used iteratively solution. simulation results which applied quarter car show efficiency proposed fields optimality robustness.